Information processing apparatus

Abstract

An information processing apparatus generates multiple combinations of sensor data inputted to a machine learning apparatus, inputs the combinations of sensor data to the machine learning apparatus, and generates a recognizer corresponding to each of the combinations of sensor data. Further, the performance of the recognizers is evaluated in accordance with expected performance required for the recognizers, and the combinations of sensor data corresponding to the recognizers satisfying the expected performance are outputted. Thus, the rates of contribution of two or more pieces of sensor data inputted to the machine learning apparatus are evaluated, and the configuration of sensors is optimized.

The invention claimed is: 
     
       1. An information processing apparatus including a machine learning apparatus configured to receive a plurality of different pieces of sensor data and perform machine learning, the information processing apparatus comprising:
 an input data control section configured to set a selection rule for generating one or more combinations of the sensor data; 
 an input data selection section configured to generate a plurality of combinations of the sensor data in accordance with the selection rule, input the combinations of the sensor data to the machine learning apparatus, and generate a recognizer corresponding to each of the combinations of the sensor data; 
 an expected-performance setting section configured to set expected performance required for the recognizers; and 
 a performance evaluation section configured to evaluate performance of the recognizers in accordance with the expected performance and output the combination of the sensor data corresponding to the recognizer satisfying the expected performance, wherein 
 the expected performance is defined using cost for acquiring the sensor data, and 
 the cost for acquiring the sensor data is calculated based on a price of a sensor for acquiring the sensor data, a degree of difficulty of installation of the sensor, or time or processing load required for acquiring or processing the sensor data. 
 
     
     
       2. The information processing apparatus according to  claim 1 , wherein the expected performance is defined using correct answer rate, precision rate, recall rate, or F value of the recognizer. 
     
     
       3. The information processing apparatus according to  claim 1 , wherein the sensor data concern at least any one of temperature, displacement, pressure, current, speed, acceleration, image, illuminance, audio, smell and length. 
     
     
       4. The information processing apparatus according to  claim 1 , wherein the performance evaluation section includes an interface configured to display a plurality of combinations of the sensor data, the combinations satisfying the expected performance, and allow a user to select a desired combination of the sensor data. 
     
     
       5. The information processing apparatus according to  claim 1 , wherein the recognizer includes a neural network. 
     
     
       6. The information processing apparatus according to  claim 5 , wherein the input data control section further sets a setting rule for a configuration of an intermediate layer of the neural network. 
     
     
       7. The information processing apparatus according to  claim 5 , wherein the input data control section has the selection rule for deleting the sensor data having a small coupling coefficient between an input layer and an intermediate layer in the neural network. 
     
     
       8. The information processing apparatus according to  claim 5 , wherein the input data control section has the selection rule for deleting the sensor data having a low rate of contribution of an intermediate layer to firing at the time of recognition in the neural network. 
     
     
       9. The information processing apparatus according to  claim 1 , wherein the sensor data are collected from one or more sensors through a communication network.